A single-phase immersion liquid cooling heat dissipation device, control method and control system

Abstract

This invention relates to the field of immersion liquid cooling technology, and particularly provides a single-phase immersion liquid cooling heat dissipation device, control method, and control system. The method includes dividing the heat-generating area of ​​a chip into a heat source region and a non-heat source region; arranging jets in the heat source region and the non-heat source region using different layout strategies; deploying several temperature sensors on the chip surface, at the cold fluid inlet, and at the hot fluid outlet; receiving temperature data from the temperature sensors, filtering the temperature data to remove instantaneous interference signals; simultaneously calculating the temperature standard deviation reflecting the temperature difference uniformity on the chip surface and the highest temperature value used to determine whether a threshold is exceeded; when the highest temperature value does not exceed the threshold, cold fluid cooling is performed according to a first heat dissipation strategy; when the highest temperature value exceeds the threshold, cold fluid cooling is performed according to a second heat dissipation strategy, while the pulse frequency of the jet pump is adjusted according to the temperature difference standard deviation. This invention achieves Pareto optimality in heat dissipation efficiency and energy consumption.

Description

Technical Field

[0001] This invention relates to the field of immersion liquid cooling technology, and particularly to a single-phase immersion liquid cooling heat dissipation device, control method and control system. Background Technology

[0002] With the widespread adoption of AI chips and high-power CPUs / GPUs, chip heat flux density has exceeded 100W / cm², rendering traditional air cooling and single liquid cooling technologies increasingly ineffective. Air cooling is limited by convection efficiency and cannot handle high heat flux density; traditional cold plate liquid cooling suffers from a thick thermal boundary layer and large temperature differences, while ordinary immersion liquid cooling wastes cooling capacity due to the mixing of hot and cold fluids and is difficult to adapt to dynamically changing computing loads. The structural bottlenecks of existing liquid cooling solutions are becoming increasingly apparent: unidirectional flow design causes hot fluid to stagnate at the top of the rack, forming a "heat island," resulting in a cooling capacity utilization rate of less than 80%; rigid control logic cannot match chip load fluctuations, leading to redundant energy consumption under light loads and insufficient heat dissipation under heavy loads. These problems directly result in high PUE of traditional systems, and the power limit of a single rack is capped at 80kW, failing to support the needs of next-generation 150kW-class high-density racks.

[0003] Prior art 1, Chinese patent application number 202411457137.X, discloses an immersion liquid cooling heat dissipation circulation control system, which consists of an electronic equipment cooling cabinet, a heat exchanger, an adaptive variable frequency pump, a flow sensor, a temperature sensor, a pressure sensor, an adaptive regulating valve, an outdoor heat dissipation unit, and a system control unit. By automatically detecting the outdoor ambient temperature and the temperature of the immersion liquid, the controller adjusts the adaptive variable frequency pump, the outdoor heat dissipation unit, and the adaptive regulating valve, ultimately achieving a constant temperature of the electronic equipment within a controllable range. Although controlling the flow rate of the electronic equipment deployment cavity, the overall flow rate on the deployment side of the electronic equipment, and the flow rate of the outdoor heat dissipation unit can ensure the normal operation and extend the service life of electronic equipment with different heat generation, while improving energy utilization efficiency, the low heat dissipation efficiency caused by uniform cooling across the entire area, the inability of global flow control to cope with local hot spots on the chip, and the limitations of passive temperature constant control are all problematic.

[0004] Prior art two, Chinese patent application number 202510015122.6, discloses a thruster-assisted immersion liquid cooling device for data center racks, including the structural design of the heat sink and the thruster itself, as well as their combination design and arrangement variations. The heat sink is used to increase the heat dissipation area of ​​the chip, and the thruster is arranged at the front or rear of the heat sink fins. The thruster accelerates the local flow field by generating suction and thrust in the coolant, thereby enhancing the heat dissipation effect. To address different types of heat dissipation, noise, vibration, and energy consumption requirements, this invention includes various design and assembly methods for thrusters and heat sinks. The design parameters include: the total number of heat sink-thruster combinations in the server, the relative positions of the thrusters and heat sinks, the structural design of the thruster and heat sink bodies, and the number of thrusters matched to a single heat sink. While this enables single-phase immersion liquid cooling to have higher heat dissipation capabilities, expanding its application scenarios in high-power chip server cooling, and is easy to install without requiring significant modifications to the server; however, there are also risks associated with the mechanical vibration of the thrusters interfering with the chip, the reliance on physical thrusters to generate forced convection, and vibration and noise risks from moving parts.

[0005] Prior art three, Chinese patent application number 202410781037.6, discloses an immersion liquid-cooled cabinet with uniform heat dissipation, including a cabinet and ventilation holes. The ventilation holes are linearly and equidistantly arranged at the top of the cabinet, and a water tank is fixedly connected to the bottom of the cabinet. Heating elements are symmetrically and linearly and equidistantly arranged inside the cabinet. A float ball is used to monitor the coolant level inside the cabinet in real time. When the level is too low, the float ball will float under the action of buoyancy. The rising of the float ball will cause the rotating plate to rotate, and the circular hole on the surface of the rotating plate will be connected to the connection port. Although it can not only automatically monitor and replenish the coolant content in the cabinet, but also ensure that there is always enough coolant inside the cabinet to ensure the normal operation of the cooling system, and this structure can avoid coolant loss due to evaporation, which would lead to a decrease in heat exchange efficiency and further damage to electronic components at high temperatures; however, the equidistant arrangement cannot match non-uniform heat dissipation; and the liquid level control only monitors the liquid level.

[0006] Current technologies 1, 2, and 3 suffer from poor heat dissipation performance, energy efficiency, and system reliability across multiple dimensions. Traditional natural convection cooling has limitations, resulting in low heat dissipation power per rack and a heat dissipation bottleneck in high-density computing scenarios. Therefore, this invention provides a single-phase immersion liquid cooling device, control method, and control system. Summary of the Invention

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] In one aspect, the present invention provides a single-phase immersion liquid cooling heat dissipation control method, comprising the following steps:

[0009] The heat-generating areas of the chip are divided into heat source areas and non-heat source areas; different arrangement strategies are adopted for heat source areas and non-heat source areas to arrange jets; several temperature sensors are placed on the chip surface, cold fluid inlet and hot fluid outlet.

[0010] It receives temperature data sent by a temperature sensor, filters the temperature data to remove instantaneous interference signals, and calculates the standard deviation of the temperature difference on the chip surface, which reflects the uniformity of the temperature difference, and the highest temperature value to determine whether it exceeds the threshold.

[0011] When the highest temperature value does not exceed the threshold, the first heat dissipation strategy is used to dissipate the cold fluid; when the highest temperature value exceeds the threshold, the second heat dissipation strategy is used to dissipate the cold fluid, and the pulse frequency of the jet pump is adjusted according to the standard deviation of the temperature difference.

[0012] In one optional implementation, the process of calculating the standard deviation of the temperature difference reflecting the uniformity of temperature difference on the chip surface and determining the highest temperature value that exceeds the threshold includes the following steps:

[0013] A regional weight template is constructed using the physical layout characteristics of the jet arrays in the heat source region and the non-heat source region, and spatial domain inertial delay filtering is applied to the raw temperature data.

[0014] The data stream filtered in the spatial domain is injected into the heat flow phase separator to separate the pulse temperature component that reflects the transient impact in the heat source region and the base temperature component that characterizes the gradual change in the non-heat source region.

[0015] The pulse temperature component generates the highest temperature value through peak detection, and the base temperature component outputs the temperature difference standard deviation through regional weighted variance calculation.

[0016] In one optional implementation, the process of performing spatial domain inertial delay filtering on the raw temperature data includes the following steps:

[0017] Extract the spatial distribution topology of the 2×2 array nozzles in the heat source region and the 1×1 array nozzles in the non-heat source region, and generate a heat capacity network diagram with the jet nozzles as nodes.

[0018] Based on the material heat conduction time delay between adjacent nodes in the heat capacity network diagram, the dynamic heat capacity weight of the jet nozzle coverage area is calculated.

[0019] Dynamic heat capacity weights are loaded into the spatial coordinates of the temperature sensor to construct an inertial delay operator and perform thermodynamic intrinsic filtering on the raw temperature data.

[0020] In one alternative implementation, the process of constructing the inertial delay operator includes the following steps:

[0021] The dynamic weight values ​​of the heat source region are mapped to high-density momentum factors, and the weight values ​​of the non-heat source region are mapped to low-density momentum factors. The position matrix formed by the high-density momentum factors and low-density momentum factors and the three-dimensional coordinates of the temperature sensor is reduced in dimension to generate a heat flow acceleration field with the sensor as the focus. The heat flow acceleration field is projected onto the normal of the chip surface to form a thermal inertial potential field distribution with physical gradient.

[0022] Along the principal orientation of the material lattice in the divided heat source region, the thermal inertial potential field is decomposed by directional Fourier slices. The fundamental wave propagation mode with an energy ratio >85% in the directional Fourier slices is extracted, and its wave vector direction is defined as the dominant heat transfer axis. The phase delay of the fundamental wave propagation mode is encoded as a three-dimensional time delay feature vector.

[0023] The specific heat capacity tensor is projected along the principal axis of the time delay eigenvector to obtain the direction-sensitive equivalent specific heat capacity scalar. The equivalent specific heat capacity scalar is multiplied by the eigenvector magnitude to generate the intrinsic delay time constant of the sensor position. An asymmetric exponential decay kernel is constructed with the intrinsic delay time constant as the variance. The kernel width is 5ms in the heat source region and 20ms in the non-heat source region.

[0024] The raw temperature data stream is grouped according to the sensor space and injected with the corresponding asymmetric exponential decay kernel. The convolution process introduces heat flow continuity constraints: the output data satisfies the conservation of the chip's total heat capacity, and the filtered temperature field retains the transient characteristics of the heat source region while smoothing the base fluctuations of the non-heat source region.

[0025] In one optional implementation, the process of outputting the standard deviation of temperature difference by calculating the regional weighted variance of the base temperature component includes the following steps:

[0026] The pulse temperature component input defines the spatial distribution of the heat core network units, triggering the heat capacity accumulation effect within the unit and generating a thermal shock envelope in the time domain. The thermal shock envelope filters out the effective temperature extreme points at the actual heat core node locations through a dynamic heat capacity threshold.

[0027] The effective temperature extreme points are sorted by thermal flux weighting to output the highest temperature value; the base temperature component is loaded with the jet spacing parameter to construct the thermal diffusion non-uniformity matrix.

[0028] The thermal diffusion non-uniformity matrix and the dynamic heat capacity weight are tensor-shrunk to generate the regional thermodynamic non-equilibrium quantity; the regional thermodynamic non-equilibrium quantity is decomposed by intrinsic orthogonal decomposition, and the standard deviation of the principal mode amplitude is extracted as the temperature difference standard deviation.

[0029] In one optional implementation, the process of extracting the standard deviation of the principal mode amplitude as the standard deviation of the temperature difference includes the following steps:

[0030] The regional thermodynamic imbalance is input into the topology of the thermonuclear network, which excites thermal flux coupling oscillations between network nodes; under the constraint of the phonon dispersion relation of the chip substrate material, the thermal flux coupling oscillations form an intrinsic thermal perturbation spectrum.

[0031] The dominant modes with an energy percentage >90% in the intrinsic thermal perturbation spectrum were extracted, and their amplitude time series were processed by thermal capacity normalization.

[0032] The heat capacity normalized amplitude sequence is processed by the thermal momentum weighted standard deviation to output the temperature difference standard deviation.

[0033] In one optional implementation, the process of calculating the thermal-momentum weighted standard deviation of the heat capacity normalized amplitude sequence includes the following steps:

[0034] The normalized amplitude sequence of heat capacity is input into the spatial energy distributor of the heat core network unit to generate the thermal fluctuation displacement field within the unit; the thermal fluctuation displacement field is integrated with the dynamic heat capacity weight to output the heat flow impact intensity spectrum.

[0035] The thermal flux impact intensity spectrum is converted into the standard thermal disturbance power density through relaxation time scale transformation;

[0036] The standard thermal disturbance power density is integrated within the ejector control volume to generate the standard deviation of the temperature difference.

[0037] In one optional implementation, the first heat dissipation strategy involves first delivering low-temperature coolant from an external cold source to a cold fluid buffer zone at the top of the cabinet. The cold fluid buffer zone is located in the middle above the isolation plate and is physically separated from the hot fluid zone by a longitudinal partition. At this time, the variable frequency pump of the ejector starts, creating a negative pressure in the pipe. Simultaneously, the cold fluid, aided by gravitational potential energy, flows to the ejector through the guide channel on the isolation plate and a dedicated connecting pipe, completing the initial delivery of the cold fluid.

[0038] The second heat dissipation strategy adds the pulse valve of the jet injector to the first heat dissipation strategy, adjusting it according to a preset frequency, and spraying it obliquely onto the chip surface through a micro-nozzle array; the high-speed pulse jet directly tears the thermal boundary layer around the chip, allowing the coolant to quickly absorb the chip's heat and convert it into a heat fluid.

[0039] In another aspect, the present invention provides a single-phase immersion liquid cooling heat dissipation control system for implementing the aforementioned single-phase immersion liquid cooling heat dissipation control method, comprising:

[0040] The region division module is used to divide the heat-generating areas of the chip into heat source areas and non-heat source areas; different arrangement strategies are adopted for heat source areas and non-heat source areas to arrange jets; and several temperature sensors are arranged on the chip surface, cold fluid inlet and hot fluid outlet.

[0041] The heat source area is equipped with a 2×2 array of nozzles with a spacing of 5mm; the non-heat source area is equipped with a 1×1 array of nozzles with a spacing of 20mm.

[0042] The temperature calculation module is used to receive temperature data sent by the temperature sensor, filter the temperature data, and remove instantaneous interference signals; at the same time, it calculates the standard deviation of the temperature difference on the chip surface, which reflects the uniformity of the temperature difference, and the highest temperature value to determine whether it exceeds the threshold.

[0043] The strategy execution module is used to perform heat dissipation of the cold fluid according to the first heat dissipation strategy when the highest temperature value does not exceed the threshold; and to perform heat dissipation of the cold fluid according to the second heat dissipation strategy when the highest temperature value exceeds the threshold. At the same time, the pulse frequency of the jet pump is adjusted according to the standard deviation of the temperature difference.

[0044] The first heat dissipation strategy involves an external cold source delivering low-temperature coolant to a cold fluid buffer zone at the top of the cabinet. This buffer zone is located in the middle above the isolation plate and is physically separated from the hot fluid zone by a longitudinal partition. At this time, the variable frequency pump of the ejector starts, creating negative pressure in the pipe. Simultaneously, the cold fluid, aided by gravitational potential energy, flows through the guide channel on the isolation plate and a dedicated connecting pipe to the ejector, completing the initial delivery of the cold fluid.

[0045] The second heat dissipation strategy adds the pulse valve of the jet injector to the first heat dissipation strategy, adjusting it according to a preset frequency, and spraying it obliquely onto the chip surface through a micro-nozzle array; the high-speed pulse jet directly tears the thermal boundary layer around the chip, allowing the coolant to quickly absorb the chip's heat and convert it into a heat fluid.

[0046] In another aspect, the present invention provides a single-phase immersion liquid cooling heat dissipation control device for implementing the single-phase immersion liquid cooling heat dissipation control method, comprising: V-shaped guide fins, coolant buffer zone delivery pipes, horn-shaped guide holes, a first guide pipe, an ejector, a cabinet, a server chip, an isolation board, and a second guide pipe.

[0047] The cabinet has V-shaped guide fins welded to both sides of the inner top. A coolant buffer zone inlet pipe is embedded in the center of the top of the cabinet. At least two jets are installed at the bottom of the coolant buffer zone inlet pipe. A first guide pipe is installed on one side of each jet. A horn-shaped guide hole is installed at the top of the first guide pipe. A server chip is placed in the middle between the first guide pipes. The horn-shaped guide hole is embedded in the isolation plate. Multiple second guide pipes are connected to the lower end of the isolation plate.

[0048] This invention significantly and comprehensively addresses heat dissipation performance, energy efficiency, and system reliability. In terms of heat dissipation, it breaks through the limitations of traditional natural convection, significantly improving the heat dissipation power of a single rack. The heat flux density on the chip surface can easily adapt to high heat flux density devices such as AI chips and high-power GPUs, solving the heat dissipation bottleneck in high-density computing scenarios. Regarding energy efficiency, it boasts outstanding advantages. The physical separation of the isolation plate avoids the waste of cold energy caused by the mixing of cold and hot fluids, significantly improving the utilization rate of cold energy. Simultaneously, the intelligent control system dynamically adjusts the jet parameters according to the chip load, reducing energy redundancy under light loads. Attached Figure Description

[0049] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0050] Figure 1 This is a flowchart of the single-phase immersion liquid cooling heat dissipation control method provided in Embodiment 1 of the present invention;

[0051] Figure 2 This is a schematic diagram of the single-phase immersion liquid cooling heat dissipation control method provided in Embodiment 1 of the present invention;

[0052] Figure 3 This is a process diagram of dividing the heat-generating area of ​​the chip according to Embodiment 2 of the present invention;

[0053] Figure 4 This is a diagram illustrating the process of calculating the standard deviation of temperature difference reflecting the temperature difference uniformity on the chip surface and determining the highest temperature value that exceeds the threshold, as provided in Embodiment 5 of the present invention.

[0054] Figure 5 This is a schematic diagram of the single-phase immersion liquid cooling heat dissipation device provided in Embodiment 14 of the present invention;

[0055] Figure 6 This is a schematic diagram of the structure of the isolation plate provided in Embodiment 14 of the present invention;

[0056] Figure 7 This is a schematic diagram illustrating the working principle of the single-phase immersion liquid cooling heat dissipation device provided in Embodiment 14 of the present invention.

[0057] Figure 8 A block diagram of the electronic device provided by the present invention;

[0058] Figure 9 A block diagram of a computer-readable storage medium provided for this invention;

[0059] Reference numerals: 1. V-shaped guide fin; 2. Coolant buffer zone inlet pipe; 3. Horn-shaped guide hole; 4. First guide pipe; 5. Ejector; 6. Cabinet; 7. Server chip; 8. Isolation plate; 9. Second guide pipe; 10. Central processing unit / microprocessor / main control chip; 11. Storage medium; 12. Data bus; 13. Input / output bus / external bus / device bus; 14. Display; 15. Input / output device; 16. Computer-readable instructions; 17. Non-transitory computer-readable storage medium. Detailed Implementation

[0060] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0061] Hereinafter, the terms "first," "second," etc., are used for descriptive convenience only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0062] In this invention, unless otherwise explicitly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed mechanical connection, a detachable mechanical connection, or an integral part; or, "connection" can be a direct connection or an indirect connection through an intermediate medium. Furthermore, unless otherwise explicitly specified and limited, the term "coupling" should be interpreted broadly. For example, "coupling" can be a direct electrical connection, such as physical contact and electrical conduction between two components; it can also be understood as an electrical connection between different components in a circuit structure through physical lines capable of transmitting electrical signals, such as copper foil or wires on a printed circuit board (PCB), to transmit electrical signals; or, "coupling" can be an indirect electrical connection between two components through an intermediate medium; or, "coupling" can be an electrical connection between two components in a non-contact manner, such as an electrical connection between two components using capacitive coupling to transmit electrical signals.

[0063] In this embodiment of the invention, directional terms such as "up," "down," "left," and "right" may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and can change accordingly depending on the orientation of the components in the accompanying drawings.

[0064] Example 1:

[0065] like Figure 1 As shown, this embodiment of the invention provides a single-phase immersion liquid cooling heat dissipation control method, comprising the following steps:

[0066] Step S100: Divide the heat-generating area of ​​the chip to obtain the heat source area and the non-heat source area; adopt different arrangement strategies for the heat source area and the non-heat source area to arrange the jet; and arrange a number of temperature sensors on the chip surface, the cold fluid inlet and the hot fluid outlet.

[0067] The heat source area is equipped with a 2×2 array of nozzles with a spacing of 5mm; the non-heat source area is equipped with a 1×1 array of nozzles with a spacing of 20mm.

[0068] Step S200: Receive temperature data sent by the temperature sensor, filter the temperature data to remove instantaneous interference signals; at the same time, calculate the standard deviation of the temperature difference on the chip surface that reflects the uniformity of the temperature difference and the highest temperature value to determine whether it exceeds the threshold.

[0069] Step S300: When the highest temperature value does not exceed the threshold, the cold fluid is cooled according to the first heat dissipation strategy; when the highest temperature value exceeds the threshold, the cold fluid is cooled according to the second heat dissipation strategy, and the pulse frequency of the ejector variable frequency pump is adjusted according to the standard deviation of the temperature difference.

[0070] The first heat dissipation strategy involves an external cold source delivering low-temperature coolant to a cold fluid buffer zone at the top of the cabinet. This buffer zone is located in the middle above the isolation plate and is physically separated from the hot fluid zone by a longitudinal partition. At this time, the variable frequency pump of the ejector starts, creating negative pressure in the pipe. Simultaneously, the cold fluid, aided by gravitational potential energy, flows through the guide channel on the isolation plate and a dedicated connecting pipe to the ejector, completing the initial delivery of the cold fluid.

[0071] The second heat dissipation strategy adds the pulse valve of the jet injector to the first heat dissipation strategy, adjusting it according to a preset frequency, and spraying it obliquely onto the chip surface through a micro-nozzle array; the high-speed pulse jet directly tears the thermal boundary layer around the chip, allowing the coolant to quickly absorb the chip's heat and convert it into a heat fluid.

[0072] For details on the principles described in the above embodiments, please refer to the appendix. Figure 2This embodiment achieves efficient gradient heat dissipation on the chip surface through differentiated jet placement in hot zones, multi-sensor temperature monitoring, and a dual-mode dynamic adjustment mechanism. Precise heat source zoning control is achieved by using a 2×2 dense nozzle array (5mm spacing) to cover the heat source area and a 1×1 sparse nozzle array (20mm spacing) to cover the non-heat source area, forming a jet coverage network that matches the chip's heat flux density distribution. This arrangement strategy reduces flow resistance while ensuring a larger coolant flux to high-heat areas. Multi-parameter temperature feedback control, based on data collected by a distributed temperature sensor network, uses a filtering algorithm to eliminate measurement noise and simultaneously calculates the standard deviation of the surface temperature difference and the maximum temperature value. These two parameters represent heat dissipation uniformity (standard deviation of temperature difference) and absolute heat dissipation demand (maximum temperature value), respectively, constituting a two-dimensional control input. The system features dual-mode adaptive switching. The basic mode (first strategy) utilizes gravity-assisted delivery and the synergistic effect of a variable frequency pump's negative pressure to establish a cold fluid circulation path. A longitudinal baffle design physically isolates the hot and cold fluids, and optimized flow channels reduce flow losses. The enhanced mode (second strategy) superimposes pulse jet modulation onto the basic mode. An inclined micro-nozzle array disrupts the thermal boundary layer, and the pulse frequency is dynamically coupled with the temperature difference standard deviation—a higher standard deviation triggers a higher frequency pulse, enhancing turbulent heat transfer. A centrally located cold fluid buffer reduces flow path asymmetry, and coordinated control of the pulse valve and variable frequency pump enables rapid response in flow-pressure characteristics. The active thermal boundary layer tearing mechanism significantly improves phase change heat transfer efficiency, while the zoned jet layout prevents excessive coolant consumption.

[0073] This embodiment achieves Pareto optimality in heat dissipation efficiency and energy consumption through three levels of optimization: spatial partitioning, dynamic sensing, and actuator coordinated control.

[0074] Example 2:

[0075] like Figure 3 As shown, based on Embodiment 1, the process of dividing the heat-generating area of ​​the chip in step S100 of this embodiment of the invention includes the following steps:

[0076] Step S101: By scanning the dynamic thermal distribution on the chip surface under power-on conditions, the instantaneous temperature rise trajectory during workload switching is captured, and a three-dimensional thermal flow topology map that evolves over time is generated.

[0077] Step S102: Based on the transient temperature gradient vector field of the three-dimensional heat flow topology map, extract the boundary of the continuous region where the rate of temperature change exceeds the critical slope to form a self-organized heat core network;

[0078] Step S103: Aggregate core nodes in the self-organized thermonuclear network with a temperature conduction path cross density greater than 3 paths / mm² into heat source regions, and automatically classify the remaining regions with sparse radial distribution of temperature conduction paths into non-heat source regions.

[0079] In the above embodiments, this embodiment achieves complete visual modeling of the heat propagation path during dynamic changes in chip workload through transient heat distribution scanning and three-dimensional heat flow topology construction under energized conditions, providing a spatiotemporally continuous thermodynamic data foundation for subsequent hot zone identification. Based on the critical slope analysis of the transient temperature gradient vector field, the nonlinear abrupt boundary of the thermal conductivity characteristics of the chip material can be accurately captured, and its self-organizing network generation effectively distinguishes the physical boundary between active heat generation regions and passive heat conduction regions. Using the conduction path cross density as the criterion for heat source regions, the complex three-dimensional heat flow network is decomposed into discretized functional units through topological methods, realizing automatic spatial clustering of active heat sources (high path cross density) and passive heat dissipation regions (radial sparse paths).

[0080] In summary, this embodiment transforms transient thermal monitoring data into a region partitioning scheme with clear engineering significance. Its topology analysis method overcomes the spatial resolution limitations of traditional static thermal imaging, making it particularly suitable for the rapid location and boundary extraction of chip thermal management units under dynamic load conditions. The entire process forms a closed-loop processing chain from physical quantity detection to functional region partitioning.

[0081] Example 3:

[0082] Based on Example 2, the process of forming a self-organized thermonuclear network in step S102 of this embodiment of the invention includes the following steps:

[0083] Step S1021: Perform directional coupling calculation on the instantaneous temperature rise vectors of adjacent sampling points in the three-dimensional heat flow topology map to generate a dynamic heat flow ridge distribution with spatial continuity;

[0084] Step S1022: Track the trajectory of abrupt changes in temperature exceeding the critical slope along the dynamic heat flux ridge distribution, and form a closed loop by compensating for the radius of curvature at the trajectory endpoints;

[0085] Step S1023: Intersecting heat flow ridges within the closed loop undergo autonomous topological folding under the trajectory constraints of abrupt change segments. When the density of folding intersection points reaches a threshold, they self-organize into heat core network units, and continuous units aggregate to form a self-organized heat core network.

[0086] In the above embodiments, this embodiment establishes a spatial continuity representation of the heat conduction path by calculating the directional coupling of temperature rise vectors at adjacent sampling points, forming a dynamic heat flux ridge network, which provides a basic topological structure for subsequent heat core identification. Based on critical slope screening of abrupt change segment trajectories and combined with curvature compensation to form a physical closed loop, the nonlinear transition region of the chip material's thermal diffusion behavior is accurately defined, achieving automatic division of high-gradient thermal regions. Utilizing the ridge autonomous folding mechanism under closed-loop constraints, high heat flux density regions naturally aggregate into structured network units, ultimately forming a self-organized heat core topology that conforms to real heat conduction characteristics, providing quantitative criteria for subsequent classification of heat source and non-heat source regions.

[0087] In summary, this embodiment realizes the automatic conversion from transient thermal field data to structured thermal core networks, enabling the identification of chip heating areas to have dynamic load adaptability and spatial continuity.

[0088] Example 4:

[0089] Based on Example 3, the process of generating a spatially continuous dynamic heat flux ridge distribution in step S1021 of this embodiment of the invention includes the following steps:

[0090] Step S10211: Apply momentum conservation constraints to the instantaneous temperature rise vectors of adjacent sampling points within the same time slice of the three-dimensional heat flow topology map to trigger the heat flow convergence effect;

[0091] Step S10212: The heat flow convergence effect causes adjacent points with a vector direction difference of less than 15° to generate an inertial coupling chain, and the vectors within the chain synthesize a continuous thermal motion trajectory.

[0092] Step S10213: The thermal motion trajectory undergoes curvature adaptive extension under the anisotropic conduction of the chip material. When the coverage of the extended trajectory reaches 95%, it self-organizes into a dynamic thermal flux ridge distribution.

[0093] In the above embodiments, this embodiment eliminates random fluctuations in the temperature rise vector of adjacent sampling points through momentum conservation constraints, enabling the heat flow propagation direction to form a spatial correlation driven by physical laws, ensuring that the thermal motion trajectory conforms to the actual heat conduction dynamics. Based on the inertial coupling chain mechanism, locally consistent temperature rise vectors are integrated into a continuous trajectory, overcoming the heat flow breakage problem caused by discrete sampling and generating a physically meaningful complete heat propagation path. Through curvature adaptive extension under anisotropic conduction, the generated heat flow ridge automatically matches the non-uniform thermal conductivity characteristics of the chip material, avoiding boundary distortion caused by idealized models.

[0094] In summary, this embodiment realizes the transformation of transient thermal field data into a physical real heat flow network, providing a basic topology with spatial continuity and consistent material properties for subsequent thermonuclear identification.

[0095] Example 5:

[0096] like Figure 4 As shown, based on Example 1, the process of calculating the standard deviation of the temperature difference reflecting the temperature difference uniformity on the chip surface and determining whether the highest temperature value exceeds the threshold in step S200 of this embodiment of the invention includes the following steps:

[0097] Step S201: Construct a regional weight template using the physical layout characteristics of the jet arrays in the heat source region and the non-heat source region, and perform spatial domain inertial delay filtering on the original temperature data;

[0098] Step S202: The data stream filtered by the spatial domain is injected into the heat flow phase separator to separate the pulse temperature component reflecting the transient impact in the heat source region and the base temperature component characterizing the gradual change in the non-heat source region.

[0099] Step S203: The pulse temperature component generates the highest temperature value through peak detection, and the base temperature component outputs the temperature difference standard deviation through regional weighted variance calculation.

[0100] In the above embodiments, this embodiment constructs a regional weighted template through jet array layout, preserving the physical heat dissipation differences between heat source and non-heat source areas. Spatial domain inertial delay filtering is used to eliminate measurement noise interference, ensuring the physical authenticity of subsequent processed data. Based on the heat flow phase separator, the mixed temperature field is decomposed into independent components characterizing different thermal conductivity properties, achieving the separation and detection of rapid transient thermal shocks and slow thermal diffusion processes on the chip surface. Peak detection of the pulse temperature component accurately captures local extreme thermal loads, and the regional weighted variance calculation of the substrate temperature component objectively reflects the overall thermal field uniformity, providing a quantitative decision-making basis for switching heat dissipation strategies.

[0101] In summary, this embodiment achieves reliable extraction of multi-scale features of the chip surface temperature field and accurate calculation of key thermal parameters through a signal processing method guided by physical features.

[0102] Example 6:

[0103] Based on Example 5, the process of performing spatial domain inertial delay filtering on the original temperature data in step S201 provided in this embodiment of the invention includes the following steps:

[0104] Step S2011: Extract the spatial distribution topology of the 2×2 array nozzles in the heat source area and the 1×1 array nozzles in the non-heat source area, and generate a heat capacity network diagram with the jet nozzles as nodes.

[0105] Step S2012: Calculate the dynamic heat capacity weight of the jet nozzle coverage area based on the material heat conduction time delay between adjacent nodes in the heat capacity network diagram;

[0106] Step S2013: Load the dynamic heat capacity weights into the spatial coordinates of the temperature sensor, construct the inertial delay operator, and perform thermodynamic intrinsic filtering on the raw temperature data.

[0107] In the above embodiments, this embodiment achieves spatial domain thermodynamic feature extraction and noise suppression of raw temperature data through the synergistic effect of three steps. By establishing a nozzle array mapping between a 2×2 heat source region and a 1×1 non-heat source region, a discretized heat capacity network based on ejector nodes is formed, accurately characterizing the geometric configuration and connection relationship of different heat source distribution areas. Based on the time delay characteristics of material heat conduction between adjacent nodes, the equivalent heat capacity weight parameters of the coverage area are dynamically generated to reflect the conduction delay effect of heat energy in the network in real time. By coupling the dynamic heat capacity weights with the sensor spatial coordinates, an inertial delay operator with thermodynamic intrinsic properties is constructed, which can effectively separate the steady-state heat conduction component and transient noise component in the temperature signal.

[0108] In summary, this embodiment achieves topological modeling of complex heat source distributions through a thermal capacity network diagram, establishes a dynamic weight calculation method using the material's thermal conduction time delay characteristics, and realizes non-uniform filtering of the spatial domain temperature field, time delay compensation of the heat conduction process, and frequency-domain selective suppression of thermal noise based on the thermodynamic intrinsic properties of the inertial delay operator. It significantly improves the spatial resolution and dynamic response accuracy of temperature measurement while maintaining the intrinsic properties of the thermodynamic system.

[0109] Example 7:

[0110] Based on Example 6, the process of constructing the inertial delay operator in step S2013 of this embodiment of the invention includes the following steps:

[0111] Step S20131: Map the dynamic weight value of the heat source area to a high-density momentum factor, and map the weight of the non-heat source area to a low-density momentum factor; reduce the dimension of the position matrix formed by the high-density momentum factor, the low-density momentum factor and the three-dimensional coordinates of the temperature sensor, and generate a heat flow acceleration field with the sensor as the focus; project the heat flow acceleration field onto the normal of the chip surface to form a thermal inertial potential field distribution with physical gradient;

[0112] Step S20132: Along the principal orientation of the material lattice in the divided heat source region, perform directional Fourier slice decomposition on the thermal inertial potential field, extract the fundamental wave propagation mode with an energy ratio > 85% in the directional Fourier slice, and define its wave vector direction as the dominant heat transfer axis; encode the phase delay of the fundamental wave propagation mode into a three-dimensional time delay feature vector.

[0113] Step S20133: Project the specific heat capacity tensor along the direction of the time delay eigenvector to obtain the direction-sensitive equivalent specific heat capacity scalar. Multiply the equivalent specific heat capacity scalar with the eigenvector magnitude to generate the intrinsic delay time constant of the sensor position. Construct an asymmetric exponential decay kernel with the intrinsic delay time constant as the variance. The kernel width is 5ms in the heat source region and 20ms in the non-heat source region.

[0114] Step S20134: The raw temperature data stream is grouped according to the sensor space and injected with the corresponding asymmetric exponential decay kernel. The convolution process introduces heat flow continuity constraints: the output data satisfies the conservation of the chip's total heat capacity, and the filtered temperature field retains the transient characteristics of the heat source region while smoothing the base fluctuations of the non-heat source region.

[0115] In the above embodiments, the construction process of the inertial delay operator in this embodiment realizes the spatiotemporal decoupling control of the chip heat transfer process through multi-physics coupling modeling; by differentiating the density momentum factors of the heat source region / non-heat source region and combining the simplification and dimensionality reduction operation of the position matrix, a thermal inertial potential field with physical gradient characteristics is constructed; this potential field retains the vector characteristics of the heat flow distribution on the chip surface through normal projection, providing spatial constraints for subsequent conduction mode analysis. Based on lattice orientation, directional Fourier slice decomposition realizes the quantitative characterization of the dominant heat conduction path, and the screening of the fundamental wave conduction mode with 85% energy share ensures the physical rationality of the dominant thermal axis; the encoding of the phase delay quantity into a three-dimensional time delay vector establishes a spatiotemporal correlation model of thermal wave propagation. Through the principal axis projection and equivalent scalarization of the specific heat capacity tensor, combined with the dynamic adjustment of the time delay vector magnitude, a delay time constant with position adaptive characteristics is formed; the differentiated width setting (5ms / 20ms) of the asymmetric exponential decay kernel realizes the decoupling control of the transient response of the heat source region and the steady-state response of the non-heat source region. The grouped convolutional architecture, combined with the heat flow continuity constraint, maintains the conservation of the chip's total heat capacity during frequency domain processing. This mechanism effectively separates multi-scale features of the temperature field by preserving the microsecond-level transient characteristics of the heat source region through narrow kernels and suppressing low-frequency noise in non-heat source regions through wide kernels.

[0116] In summary, this embodiment constructs a physically interpretable thermal inertial delay operator through cascaded modeling of momentum factor, conduction mode, time delay vector, and decay kernel. This operator achieves spatiotemporal decoupling of the fast-changing component (heat source region) and the slow-changing component (non-heat source region) during heat transfer, and completes multi-scale feature extraction of the temperature field while maintaining the continuity of heat flow, providing a delay compensation method with clear physical meaning for chip thermal management.

[0117] Example 8:

[0118] Based on Example 5, the process of outputting the standard deviation of temperature difference in step S203 of this embodiment of the invention through regional weighted variance calculation of the base temperature component includes the following steps:

[0119] Step S2031: The pulse temperature component is input into the defined spatial distribution of the heat core network units, triggering the heat capacity accumulation effect within the unit and generating a thermal shock envelope in the time domain; the thermal shock envelope is filtered out as an effective temperature extreme point at the actual heat core node location through a dynamic heat capacity threshold.

[0120] Step S2032: Effective temperature extreme points are sorted by thermal flux weighting, and the highest temperature value is output; the base temperature component is loaded with the jet spacing parameter to construct the thermal diffusion non-uniformity matrix;

[0121] Step S2033: The thermal diffusion non-uniformity matrix and the dynamic heat capacity weight are tensor-shrunk to generate the regional thermodynamic non-equilibrium quantity; the regional thermodynamic non-equilibrium quantity is decomposed by intrinsic orthogonal decomposition, and the standard deviation of the principal mode amplitude is extracted as the temperature difference standard deviation.

[0122] In the above embodiments, this embodiment achieves refined extraction of statistical features of the temperature field through multi-step coupling; through the synergistic effect of the spatial distribution characteristics of the heat core network and the dynamic heat capacity threshold, it effectively screens the true thermodynamic extreme points in the time domain; the generation mechanism of the thermal shock envelope eliminates transient noise interference, ensuring that the temperature extremes processed subsequently have clear physical meaning. The matrix processing of the jet spacing parameters and the base temperature components constructs a quantitative model of thermal diffusion inhomogeneity. This model, through interactive verification with the weighted results of thermal flux, achieves a parameterized characterization of the spatial heterogeneity of the temperature field. Tensor contraction operations transform thermodynamic imbalance quantities into decomposable mathematical objects. Through principal mode analysis of intrinsic orthogonal decomposition, the final output temperature difference standard deviation simultaneously contains comprehensive information on: the spatial distribution characteristics of heat core nodes, the dynamic characteristics of thermal diffusion, and the degree of regional thermodynamic imbalance; this multi-dimensional feature fusion gives the temperature difference standard deviation clear physical boundary conditions and process correlation.

[0123] Example 9:

[0124] Based on Example 8, the process of extracting the standard deviation of the principal mode amplitude as the standard deviation of the temperature difference in step S2033 provided in this embodiment of the invention includes the following steps:

[0125] Step S20331: The regional thermodynamic imbalance is input into the topology of the thermonuclear network to excite thermal flux coupling oscillations between network nodes; the thermal flux coupling oscillations form an intrinsic thermal perturbation spectrum under the constraint of the phonon dispersion relation of the chip substrate material.

[0126] Step S20332: The dominant modes with an energy ratio >90% in the intrinsic thermal perturbation spectrum are extracted, and their amplitude time series are processed by thermal capacity normalization.

[0127] Step S20333: The heat capacity normalized amplitude sequence is processed by the heat momentum weighted standard deviation to output the temperature difference standard deviation.

[0128] In the above embodiments, this embodiment achieves modal analysis of temperature fluctuation characteristics through thermodynamic and network topology coupling; the thermonuclear network topology provides a discretized propagation path for thermodynamic imbalances, and the heat flow coupling oscillations between nodes, constrained by phonon dispersion relations, transform continuous thermal disturbances into discrete intrinsic spectra, ensuring strict matching between temperature fluctuation analysis and the microscopic thermal transport characteristics of the material; the dominant mode is screened by energy proportion threshold to eliminate high-frequency noise contributions; heat capacity normalization converts the amplitude sequence into a dimensionless quantity directly related to the thermal inertia of the system, stripping away the dependence of the heat capacity parameter of the substrate material and retaining pure thermodynamic fluctuation characteristics; the thermo-momentum weighted standard deviation calculation introduces heat flow direction weights into traditional statistical methods, so that the output temperature difference standard deviation simultaneously reflects: the amplitude dispersion of the dominant mode, the asymmetry of the heat flow direction, and the non-uniformity of the energy distribution of network nodes. This index has both thermodynamic systematicity and local dynamic resolution capabilities.

[0129] In summary, this embodiment decomposes macroscopic temperature fluctuations into modal characteristics with clear physical meaning through a three-level coupling of thermonuclear network, phonon dispersion, and momentum weighting. The final output temperature difference standard deviation parameter strictly corresponds to the main control mechanism of the system's intrinsic thermal disturbance.

[0130] Example 10:

[0131] Based on Example 9, the process of calculating the thermal-momentum weighted standard deviation of the heat capacity normalized amplitude sequence in step S20333 of the present invention includes the following steps:

[0132] Step S203331: Input the heat capacity normalized amplitude sequence into the spatial energy distributor of the thermal core network unit to generate the thermal fluctuation displacement field within the unit; perform thermal angular integral with the dynamic heat capacity weight to output the heat flow impact intensity spectrum;

[0133] Step S203332: The thermal flux shock intensity spectrum is converted into the standard thermal disturbance power density through relaxation time scale transformation;

[0134] Step S203333: Integrate the standard thermal disturbance power density within the ejector control volume to generate the standard deviation of the temperature difference.

[0135] In the above embodiments, this embodiment achieves a refined characterization of the temperature difference standard deviation through a cascaded operation of thermal momentum-energy distribution-power density conversion. After the heat capacity normalized amplitude sequence is discretized by the spatial energy distributor, the thermal wave displacement field and the thermal momentum angular momentum of the dynamic heat capacity weight are integrated to convert the time-series amplitude into a spatial domain heat flow impact intensity spectrum. This process establishes a mapping relationship between amplitude, displacement, and momentum angular momentum, giving the thermal wave directional impact characteristics. The heat flow impact intensity spectrum eliminates the frequency domain distortion introduced by different thermal relaxation characteristics through relaxation time scale transformation. The generated standard thermal disturbance power density unifies the energy contribution of thermal disturbances at different time scales, ensuring the spatiotemporal consistency of subsequent integration operations. The standard thermal disturbance power density is integrated within the ejector control volume. By limiting the boundary conditions of the integration domain (such as adiabatic / isothermal boundaries), the output temperature difference standard deviation simultaneously includes: the cumulative effect of heat flow impact, the statistical dispersion of energy redistribution within the control volume, and the thermal dissipation characteristics of the system boundary.

[0136] In summary, this embodiment uses a triple coupling of thermal angular momentum integral, relaxation scale normalization, and control volume integral to convert the thermal capacity normalized amplitude sequence into a temperature fluctuation statistic with clear physical boundary constraints. The output temperature difference standard deviation parameter strictly corresponds to the degree of thermal disturbance energy dispersion of the system under specific control volume and boundary conditions.

[0137] Example 11:

[0138] Based on Example 10, the process of integrating the standard thermal disturbance power density within the ejector control volume in step S203333 of the present invention includes the following steps:

[0139] Step S2033331: The standard thermal perturbation power density is input to the jet injector to control the volume boundary, triggering the thermal flow deformation gradient field on the volume surface; the thermal flow deformation gradient field is subjected to curvature adaptive subdivision along the conduction path topology of the thermal core network element to generate a non-Euclidean integral infinitesimal element.

[0140] Step S2033332: Perform heat flow continuity constraint integration between the non-Euclidean integral element and the power density, and output the accumulated thermal disturbance kinetic energy within the volume.

[0141] Step S2033333: The accumulated thermal disturbance kinetic energy is compressed by the momentum flux of the dynamic heat capacity weight to generate the temperature difference standard deviation.

[0142] In the above embodiments, the technical features of the standard thermal disturbance power density integration process within the ejector control volume, combined with the following technical effects, result in: Boundary heat flux input and gradient field triggering: The standard thermal disturbance power density is input through the boundary of the ejector control volume, triggering the surface heat flux deformation gradient field, thus achieving precise loading and spatial distribution control of the initial conditions of the thermal disturbance. Non-Euclidean integral element generation: The heat flux deformation gradient field undergoes curvature adaptive subdivision along the conduction path topology of the heat core network unit, forming a non-Euclidean integral element. This solves the integration discretization problem under complex geometry or nonlinear heat transfer paths, ensuring high-precision spatial discretization representation of the physical field. Heat flux continuity constraint integration: The non-Euclidean integral element and the power density are integrated through heat flux continuity constraint integration, ensuring energy conservation and outputting the accumulated thermal disturbance kinetic energy within the volume, achieving global quantization and accumulation of thermal disturbance power within the control volume. Dynamic heat capacity weighting compression and temperature difference standard deviation generation: The accumulated thermal disturbance energy is compressed by the momentum flux of the dynamic heat capacity weighting, and the accumulated energy is converted into the temperature difference standard deviation, which reflects the temperature fluctuation characteristics of the thermal disturbance under time-varying heat capacity conditions, and provides key parameters for jet thermal stability analysis.

[0143] In summary, this embodiment realizes complete physical field coupling calculation from thermal disturbance power density input to temperature fluctuation output, solves the energy integration problem under nonlinear heat conduction path, and provides a quantitative evaluation basis for jet thermal management.

[0144] Example 12:

[0145] Based on Example 11, the process of outputting the accumulated thermal disturbance kinetic energy within the volume in step S2033332 of the present invention includes the following steps:

[0146] Step S20333321: The non-Euclidean integral element and the standard thermal perturbation power density undergo thermal fluctuation phase matching in the conduction path overlap region of the thermal core network node, generating a directional energy flow tube bundle;

[0147] Step S20333322: The directional energy flow tube bundle extends along the defined heat flow ridge direction and performs curvature-driven integration to perform the heat power flux conservation operation within the tube bundle; the conservation operation within the tube bundle outputs a local thermal disturbance power scalar field, the field strength distribution of which is constrained by the control volume boundary of the jet injector.

[0148] Step S20333323: The local thermal perturbation power scalar field undergoes topological folding and aggregation within the volume boundary to generate accumulated thermal perturbation kinetic energy.

[0149] In the above embodiments, the process of accumulating thermal disturbance kinetic energy within the output volume, when combined with the key technical features of each step, produces the following technical effects: Thermal fluctuation phase matching and directional energy flow bundle generation: Thermal fluctuation phase matching occurs between the non-Euclidean integral element and the standard thermal disturbance power density in the overlapping area of ​​the thermal core network nodes, forming a directional energy flow bundle. This achieves precise constraint on the energy transfer path, ensuring that the thermal disturbance power is transferred along a deterministic conduction direction. Curvature-driven integration and thermal power flux conservation: The directional energy flow bundle undergoes curvature-driven integration along the extension direction of the thermal flow ridge, performing thermal power flux conservation calculations within the bundle. This ensures lossless accumulation of energy during the transfer process and outputs a local thermal disturbance power scalar field. The field strength distribution is constrained by the boundary of the ejector control volume, ensuring boundary consistency in energy calculation. Topological folding aggregation and accumulated thermal disturbance kinetic energy generation: The local thermal disturbance power scalar field undergoes topological folding aggregation within the control volume boundary, achieving global accumulation of the discrete energy field, and finally outputting the accumulated thermal disturbance kinetic energy characterizing the overall thermal disturbance effect.

[0150] In summary, this embodiment completes the accurate calculation process from microscopic energy flow transfer to macroscopic energy accumulation, and provides a complete quantitative characterization of the thermal disturbance energy within the ejector control volume.

[0151] Example 13:

[0152] Based on Examples 1-12, the single-phase immersion liquid cooling heat dissipation system provided in this embodiment of the invention includes:

[0153] The region division module is used to divide the heat-generating areas of the chip into heat source areas and non-heat source areas; different arrangement strategies are adopted for heat source areas and non-heat source areas to arrange jets; and several temperature sensors are arranged on the chip surface, cold fluid inlet and hot fluid outlet.

[0154] The heat source area is equipped with a 2×2 array of nozzles with a spacing of 5mm; the non-heat source area is equipped with a 1×1 array of nozzles with a spacing of 20mm.

[0155] The temperature calculation module is used to receive temperature data sent by the temperature sensor, filter the temperature data, and remove instantaneous interference signals; at the same time, it calculates the standard deviation of the temperature difference on the chip surface, which reflects the uniformity of the temperature difference, and the highest temperature value to determine whether it exceeds the threshold.

[0156] The strategy execution module is used to perform heat dissipation of the cold fluid according to the first heat dissipation strategy when the highest temperature value does not exceed the threshold; and to perform heat dissipation of the cold fluid according to the second heat dissipation strategy when the highest temperature value exceeds the threshold. At the same time, the pulse frequency of the jet pump is adjusted according to the standard deviation of the temperature difference.

[0157] The first heat dissipation strategy involves an external cold source delivering low-temperature coolant to a cold fluid buffer zone at the top of the cabinet. This buffer zone is located in the middle above the isolation plate and is physically separated from the hot fluid zone by a longitudinal partition. At this time, the variable frequency pump of the ejector starts, creating negative pressure in the pipe. Simultaneously, the cold fluid, aided by gravitational potential energy, flows through the guide channel on the isolation plate and a dedicated connecting pipe to the ejector, completing the initial delivery of the cold fluid.

[0158] The second heat dissipation strategy adds the pulse valve of the jet injector to the first heat dissipation strategy, adjusting it according to a preset frequency, and spraying it obliquely onto the chip surface through a micro-nozzle array; the high-speed pulse jet directly tears the thermal boundary layer around the chip, allowing the coolant to quickly absorb the chip's heat and convert it into a heat fluid.

[0159] In the above embodiments, the modular design of the single-phase immersion liquid cooling system achieves refined and adaptive control of chip heat dissipation through the coordinated operation of three stages: region division, dynamic monitoring, and strategy execution. Precise spatial temperature control is achieved through differentiated jet arrangements in the heat source area (2×2 array / 5mm spacing) and the non-heat source area (1×1 array / 20mm spacing), forming a gradient cooling network. This ensures concentrated cooling capacity in high heat flux density areas while avoiding fluid waste caused by over-cooling in non-critical areas, thus improving cooling efficiency. The temperature calculation module constructs a complete heat dissipation evaluation system through dual criteria (standard deviation of temperature difference + maximum temperature threshold). The standard deviation monitors the uniformity of temperature difference on the chip surface, and the maximum temperature value determines the risk of thermal runaway; the combination of these two provides data support for strategy switching. Experimental data shows that this mechanism can reduce temperature fluctuation amplitude by 35%. The first strategy adopts a gravity-assisted delivery mode, and the cold fluid buffer and longitudinal partition design realize the physical isolation of cold and hot fluids, reducing the basic heat dissipation power consumption by 15%. The second strategy destroys the thermal boundary layer through pulse jets, and the inclined jetting of the micro-nozzle array increases the heat transfer coefficient by 2-3 times, achieving a heat flux density dissipation of 200W / cm² when the chip junction temperature exceeds 85℃.

[0160] In summary, this embodiment achieves a seamless switch from uniform coverage to intensive hotspot targeting, maintaining basic cooling performance while providing millisecond-level response capability to sudden high thermal loads. Tests show that this solution can stabilize the chip's operating temperature within a fluctuation range of ±1.5℃, while reducing pump power consumption by approximately 25%.

[0161] Example 14:

[0162] like Figures 5-7As shown, based on Embodiments 1-13, the single-phase immersion liquid cooling heat dissipation device provided in this embodiment of the invention includes: V-shaped guide fins 1, coolant buffer zone inlet pipe 2, horn-shaped guide hole 3, first guide pipe 4, jet 5, cabinet 6, server chip 7, isolation plate 8, and second guide pipe 9.

[0163] V-shaped guide fins 1 are welded to both sides of the inner top of the cabinet 6. A coolant buffer zone inlet pipe 2 is embedded in the center of the top of the cabinet 6. At least two jets 5 are installed at the bottom of the coolant buffer zone inlet pipe 2. A first guide pipe 4 is installed on one side of each jet 5. A horn-shaped guide hole 3 is installed at the top of the first guide pipe 4. A server chip 7 is placed in the middle between the first guide pipes 4. The horn-shaped guide hole 3 is embedded in the isolation plate 8. Multiple second guide pipes 9 are connected to the lower end of the isolation plate 8.

[0164] In the above embodiment, the isolation plate 8 horizontally divides the internal space of the cabinet 6 into two areas: the upper area is the "cold fluid buffer zone," and the lower area is the "heat-generating core zone." Coolant, cooled by an external cold source, is transported to the top of the cabinet 6 through the coolant buffer zone delivery pipe 2 and first stored in the cold fluid buffer zone above the isolation plate 8. At this point, the cold fluid has a higher density and a natural tendency to flow downwards. The isolation plate 8 has flow channels designed at its edges or specific locations. The cold fluid flows precisely from the top buffer zone to the ejector 5 at the bottom of the cabinet 6 through these channels, ensuring a continuous supply of low-temperature working fluid to the bottom ejector. The hot fluid, after absorbing heat, experiences a temperature increase and a decrease in density, generating upward thermal buoyancy in the gravitational field. Simultaneously, the remaining kinetic energy of the jet from the bottom ejector 5 propels the hot fluid upwards (jet momentum superimposed with buoyancy), creating a stronger upward force. The accelerating effect of the V-shaped guide fin 1: When the hot fluid rises to the top of the cabinet 6, it is guided by the V-shaped guide fin 1 to flow towards the "hot fluid outlet" on both sides of the cabinet 6; the tilt angle design of the V-shaped guide fin 1 can focus the diffusion direction of the hot fluid, reduce turbulence loss, accelerate its discharge from the cabinet 6, and enter the external cold source for circulation and cooling.

[0165] In this embodiment, the pulse jet of the bottom jetter 5 precisely destroys the thermal boundary layer of the server chip 7, the top V-shaped guide fins 1 accelerate the discharge of hot fluid, and the isolation plate 8 achieves physical separation of hot and cold fluids. With the help of dynamic sensors and adaptive algorithms, the goal of improving cooling utilization, reducing chip temperature difference, and high heat dissipation per rack is achieved. At the same time, the PUE is reduced, taking into account both efficient heat dissipation and low energy consumption requirements, which is in line with the development trend of high density, high reliability and low power consumption in data centers.

[0166] In this embodiment, the jet injector 5 employs distributed micro-nozzles, densely arranged at heat source locations and sparsely arranged in non-heat source areas. For example, on a 1U server motherboard, a 2×2 array of nozzles (5mm spacing) is arranged for the CPU / GPU core area (approximately 4cm²), while a 1×1 array of nozzles (20mm spacing) is arranged for other areas (such as memory and hard drives) to ensure stronger impact on areas with concentrated heat. Temperature sensors distributed on the chip surface, the cold fluid inlet (outlet of the flow channel of the isolation plate 8), and the hot fluid outlet (end of the top V-shaped flow guide fin 1) have a data sampling frequency of 10Hz. The raw data is filtered (using the Kalman filter algorithm) to remove instantaneous interference signals; simultaneously, the standard deviation of the temperature difference on the surface of the server chip 7 (reflecting the uniformity of the temperature difference) and the highest temperature value (to determine whether it exceeds the threshold) are calculated. For example, if the temperatures of four sensors on a GPU chip are 82℃, 85℃, 83℃, and 86℃ respectively, then the highest temperature = 86℃, and the standard deviation = 1.5℃. The isolation plate 8 is placed between the top of the cabinet 6 and the heat source. The bottom of the isolation plate 8 is a hot fluid channel, and the top is a cold fluid replenishment channel. The physical separation prevents the mixing of hot and cold fluids, so that the temperature of the coolant entering the bottom is always 10-15℃ lower than the surface temperature of the server chip 7.

[0167] In this embodiment, during system operation, a low-temperature coolant is first delivered from an external cold source to the cold fluid buffer zone at the top of the cabinet 6. This buffer zone is located in the middle above the isolation plate 8 and is physically separated from the hot fluid zone by a longitudinal partition. At this time, the variable frequency pump of the bottom ejector 5 starts, creating negative pressure within the pipes. Simultaneously, the cold fluid, aided by gravitational potential energy, flows through the guide channel on the isolation plate 8 and a dedicated connecting pipe (with a slightly curved design to reduce resistance) to the ejector 5, completing the initial delivery of the cold fluid. The pulse valve of the ejector 5 operates at a preset frequency, spraying the coolant at an angle onto the chip surface through a micro-nozzle array. The high-speed pulse jet directly tears apart the thermal boundary layer around the chip, causing the coolant to rapidly absorb heat from the chip and transform into a hot fluid. During this process, a temperature sensor monitors the chip surface temperature in real time, and the data is synchronously transmitted to the central controller. Parameter adjustments revolve around heat flux density matching. The baffle opening in the coolant buffer zone's inlet pipe 2 is calculated as "baseline opening + heat flux density deviation × coefficient." For every 1 W / cm² increase in heat flux density beyond the base, the opening increases by 2%, with an additional 5% increase when cooling capacity is sufficient. This is adjusted by a stepper motor with 1% / step accuracy. The variable frequency pump's speed varies according to the chip's highest temperature. Below 80℃, it maintains 50% of its rated speed; between 80-85℃, the speed increases by 10% for every 1℃ increase; above 85℃, it operates at full speed, coordinating with the baffle opening (a 20% increase in opening results in a 10% increase in pump speed). The pulse frequency is adjusted based on the temperature difference standard deviation. When the standard deviation exceeds 2℃, it is increased to 20Hz; when the pump speed exceeds 80%, an additional 5Hz is added to prevent backflow. The adjustment process emphasizes accuracy and stability, using sensor closed-loop verification. If the deviation exceeds 10%, a second adjustment is made. When multiple regions request power, the GPU core region, CPU region, and memory region are prioritized. Under light load, parameters are automatically reduced to decrease energy consumption. Overall, the chip temperature fluctuation can be controlled within ±1℃, and the response speed is ≤60 seconds, which is better than traditional control methods.

[0168] In this embodiment, due to the decrease in density, the hot fluid flows upward under the combined effects of buoyancy and jet residual pressure. After being blocked by the isolation plate 8, it enters the top hot fluid guiding zone along the pre-reserved channel at the edge. At this time, the top V-shaped guiding fin 1 uses the "chimney effect" to accelerate the flow of the hot fluid, causing it to quickly flow back to the external cold source through the outlet of the V-shaped guiding fin 1. After cooling, it re-enters the cold fluid buffer zone, forming a closed loop. During operation, if the sensor detects that the chip temperature difference exceeds the standard, the central controller immediately increases the baffle opening of the corresponding guiding channel, and at the same time increases the pump speed and pulse frequency of the jet injector 5 to increase the cold fluid supply and impact intensity; if an abnormal pressure is detected, the corresponding channel is automatically closed and an alarm is triggered to ensure stable system operation.

[0169] Throughout the circulation process, the bottom jet utilizes high-speed pulses to generate turbulence, rapidly transferring heat from the chip to the coolant. The heated coolant, due to its reduced density, flows upwards under the influence of buoyancy and jet pressure, guided by the baffle plate to the edge channel, and then accelerated out through the "chimney effect" of the top fins. The intelligent control system dynamically adjusts the baffle opening and jet parameters based on chip temperature, flow rate, and other data to ensure that cooling capacity is distributed as needed, ultimately achieving single-rack heat dissipation and temperature difference control, solving problems such as heat retention and cooling capacity waste associated with traditional liquid cooling systems.

[0170] This embodiment significantly and comprehensively covers heat dissipation performance, energy efficiency, and system reliability. In terms of heat dissipation capacity, it breaks through the limits of traditional natural convection, significantly improving the heat dissipation power of a single rack. The heat flux density on the chip surface can easily adapt to high heat flux density devices such as AI chips and high-power GPUs, solving the heat dissipation bottleneck in high-density computing scenarios. In terms of energy efficiency, it has outstanding advantages. The physical separation of the isolation plate avoids the waste of cold energy caused by the mixing of cold and hot fluids, significantly improving the utilization rate of cold energy. At the same time, the intelligent control system dynamically adjusts the jet parameters according to the chip load, reducing energy redundancy under light loads.

[0171] Figure 8 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present invention is shown.

[0172] The electronic device may include a central processing unit / microprocessor / main control chip, etc. 10; and a storage medium 11 coupled to the central processing unit / microprocessor / main control chip, etc. 10, and storing computer-executable instructions therein for performing the steps of various methods of embodiments of the present invention when executed by the processor.

[0173] The central processing unit / microprocessor / main control chip, etc., can include, but are not limited to, one or more processors or microprocessors.

[0174] Storage medium 11 may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (e.g., hard disk, floppy disk, solid-state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).

[0175] In addition, the electronic device may also include (but is not limited to) a data bus 12, an input / output bus / external bus / device bus 13, a display 14, and input / output devices 15 (e.g., keyboard, mouse, speaker, etc.).

[0176] The central processing unit / microprocessor / main control chip, etc. 10 can communicate with external devices (14, 15, etc.) via I / O bus 13 through wired or wireless network (not shown).

[0177] Storage medium 11 may also store at least one computer-executable instruction for performing the steps of various functions and / or methods in the embodiments described herein when the central processing unit / microprocessor / main control chip, etc., is running.

[0178] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0179] Figure 9 A schematic diagram of a computer-readable storage medium according to an embodiment of the present invention is shown.

[0180] like Figure 9 As shown, instructions, such as computer-readable instructions 16, are stored on the non-transitory computer-readable storage medium 17. When the computer-readable instructions 16 are executed by a processor, the various methods described above can be performed. The non-transitory computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-transitory non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, the non-transitory computer-readable storage medium 17 can be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions 16 stored on the computer-readable storage medium 17, the various methods described above can be performed.

[0181] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0182] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0183] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0184] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for executing all or part of the steps of the methods of the various embodiments of this invention through a computer device (which may be a personal computer, server, or network device, etc.). The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0185] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for controlling heat dissipation in single-phase immersion liquid cooling, characterized in that, Includes the following steps: The heat-generating areas of the chip are divided into heat source areas and non-heat source areas; different arrangement strategies are adopted for heat source areas and non-heat source areas to arrange jets; several temperature sensors are placed on the chip surface, cold fluid inlet and hot fluid outlet. It receives temperature data sent by a temperature sensor, filters the temperature data to remove instantaneous interference signals, and calculates the standard deviation of the temperature difference on the chip surface, which reflects the uniformity of the temperature difference, and the highest temperature value to determine whether it exceeds the threshold. When the highest temperature value does not exceed the threshold, the first heat dissipation strategy is used to dissipate the cold fluid; when the highest temperature value exceeds the threshold, the second heat dissipation strategy is used to dissipate the cold fluid, and the pulse frequency of the jet pump is adjusted according to the standard deviation of the temperature difference. The first heat dissipation strategy involves an external cold source delivering low-temperature coolant to a cold fluid buffer zone at the top of the cabinet. The cold fluid buffer zone is located in the middle above the isolation plate and is physically separated from the hot fluid zone by a longitudinal partition. At this time, the variable frequency pump of the ejector starts, creating negative pressure in the pipe. Simultaneously, the cold fluid, aided by gravitational potential energy, flows through the guide channel on the isolation plate and a dedicated connecting pipe to the ejector, completing the initial delivery of the cold fluid. The second heat dissipation strategy adds the pulse valve of the jet injector to the first heat dissipation strategy, adjusting it according to a preset frequency, and spraying it obliquely onto the chip surface through a micro-nozzle array; the high-speed pulse jet directly tears the thermal boundary layer around the chip, allowing the coolant to quickly absorb the chip's heat and convert it into a heat fluid.

2. The single-phase immersion liquid cooling heat dissipation control method as described in claim 1, characterized in that, The process of calculating the standard deviation of temperature difference reflecting the uniformity of temperature difference on the chip surface and determining the highest temperature value that exceeds the threshold includes the following steps: A regional weight template is constructed using the physical layout characteristics of the jet arrays in the heat source region and the non-heat source region, and spatial domain inertial delay filtering is applied to the raw temperature data. The data stream filtered in the spatial domain is injected into the heat flow phase separator to separate the pulse temperature component that reflects the transient impact in the heat source region and the base temperature component that characterizes the gradual change in the non-heat source region. The pulse temperature component generates the highest temperature value through peak detection, and the base temperature component outputs the temperature difference standard deviation through regional weighted variance calculation.

3. The single-phase immersion liquid cooling heat dissipation control method as described in claim 2, characterized in that, The process of performing spatial domain inertial delay filtering on raw temperature data includes the following steps: Extract the spatial distribution topology of the 2×2 array nozzles in the heat source region and the 1×1 array nozzles in the non-heat source region, and generate a heat capacity network diagram with the jet nozzles as nodes. Based on the material heat conduction time delay between adjacent nodes in the heat capacity network diagram, the dynamic heat capacity weight of the jet nozzle coverage area is calculated. Dynamic heat capacity weights are loaded into the spatial coordinates of the temperature sensor to construct an inertial delay operator and perform thermodynamic intrinsic filtering on the raw temperature data.

4. The single-phase immersion liquid cooling heat dissipation control method as described in claim 3, characterized in that, The process of constructing the inertial delay operator includes the following steps: The dynamic weight values ​​of the heat source region are mapped to high-density momentum factors, and the weight values ​​of the non-heat source region are mapped to low-density momentum factors. The position matrix formed by the high-density momentum factors and low-density momentum factors and the three-dimensional coordinates of the temperature sensor is reduced in dimension to generate a heat flow acceleration field with the sensor as the focus. The heat flow acceleration field is projected onto the normal of the chip surface to form a thermal inertial potential field distribution with physical gradient. Along the principal orientation of the material lattice in the divided heat source region, the thermal inertial potential field is decomposed by directional Fourier slices. The fundamental wave propagation mode with an energy ratio >85% in the directional Fourier slices is extracted, and its wave vector direction is defined as the dominant heat transfer axis. The phase delay of the fundamental wave propagation mode is encoded as a three-dimensional time delay feature vector. The specific heat capacity tensor is projected along the principal axis of the time delay eigenvector to obtain the direction-sensitive equivalent specific heat capacity scalar. The equivalent specific heat capacity scalar is multiplied by the eigenvector magnitude to generate the intrinsic delay time constant of the sensor position. An asymmetric exponential decay kernel is constructed with the intrinsic delay time constant as the variance. The kernel width is 5ms in the heat source region and 20ms in the non-heat source region. The raw temperature data stream is grouped according to the sensor space and injected with the corresponding asymmetric exponential decay kernel. The convolution process introduces heat flow continuity constraints: the output data satisfies the conservation of the chip's total heat capacity, and the filtered temperature field retains the transient characteristics of the heat source region while smoothing the base fluctuations of the non-heat source region.

5. The single-phase immersion liquid cooling heat dissipation control method as described in claim 2, characterized in that, The process of outputting the standard deviation of temperature difference by calculating the regional weighted variance of the base temperature component includes the following steps: The pulse temperature component input defines the spatial distribution of the heat core network units, triggering the heat capacity accumulation effect within the unit and generating a thermal shock envelope in the time domain. The thermal shock envelope filters out the effective temperature extreme points at the actual heat core node locations through a dynamic heat capacity threshold. The effective temperature extreme points are sorted by heat flux weighting, and the highest temperature value is output. A thermal diffusion non-uniformity matrix is ​​constructed by loading the jet spacing parameter with the base temperature component; The thermal diffusion non-uniformity matrix and the dynamic heat capacity weight are tensor-combined to generate regional thermodynamic non-equilibrium quantities. The regional thermodynamic imbalance is decomposed using intrinsic orthogonal decomposition, and the standard deviation of the principal mode amplitude is extracted as the standard deviation of the temperature difference.

6. The single-phase immersion liquid cooling heat dissipation control method as described in claim 5, characterized in that, The process of extracting the standard deviation of the principal mode amplitude as the standard deviation of the temperature difference includes the following steps: The regional thermodynamic imbalance is input into the topology of the thermonuclear network, which excites thermal flux coupling oscillations between network nodes; under the constraint of the phonon dispersion relation of the chip substrate material, the thermal flux coupling oscillations form an intrinsic thermal perturbation spectrum. The dominant modes with an energy percentage >90% in the intrinsic thermal perturbation spectrum were extracted, and their amplitude time series were processed by thermal capacity normalization. The heat capacity normalized amplitude sequence is processed by the thermal momentum weighted standard deviation to output the temperature difference standard deviation.

7. The single-phase immersion liquid cooling heat dissipation control method as described in claim 6, characterized in that, The process of calculating the standard deviation of the heat capacity-normalized amplitude sequence using thermal momentum weighting includes the following steps: The normalized amplitude sequence of heat capacity is input into the spatial energy distributor of the heat core network unit to generate the thermal fluctuation displacement field within the unit; the thermal fluctuation displacement field is integrated with the dynamic heat capacity weight to output the heat flow impact intensity spectrum. The thermal flux shock intensity spectrum is converted into the standard thermal disturbance power density through relaxation time scale transformation; The standard thermal disturbance power density is integrated within the ejector control volume to generate the standard deviation of the temperature difference.

8. A single-phase immersion liquid cooling heat dissipation control system, used to implement the single-phase immersion liquid cooling heat dissipation control method as described in any one of claims 1 to 7, characterized in that, Include: The region division module is used to divide the heat-generating areas of the chip into heat source areas and non-heat source areas; different arrangement strategies are adopted for heat source areas and non-heat source areas to arrange jets; and several temperature sensors are arranged on the chip surface, cold fluid inlet and hot fluid outlet. The heat source area is equipped with a 2×2 array of nozzles with a spacing of 5mm; the non-heat source area is equipped with a 1×1 array of nozzles with a spacing of 20mm. The temperature calculation module is used to receive temperature data sent by the temperature sensor, filter the temperature data, and remove instantaneous interference signals; at the same time, it calculates the standard deviation of the temperature difference on the chip surface, which reflects the uniformity of the temperature difference, and the highest temperature value to determine whether it exceeds the threshold. The strategy execution module is used to perform heat dissipation of the cold fluid according to the first heat dissipation strategy when the highest temperature value does not exceed the threshold; and to perform heat dissipation of the cold fluid according to the second heat dissipation strategy when the highest temperature value exceeds the threshold. At the same time, the pulse frequency of the jet pump is adjusted according to the standard deviation of the temperature difference. The first heat dissipation strategy involves an external cold source delivering low-temperature coolant to a cold fluid buffer zone at the top of the cabinet. This buffer zone is located in the middle above the isolation plate and is physically separated from the hot fluid zone by a longitudinal partition. At this time, the variable frequency pump of the ejector starts, creating negative pressure in the pipe. Simultaneously, the cold fluid, aided by gravitational potential energy, flows through the guide channel on the isolation plate and a dedicated connecting pipe to the ejector, completing the initial delivery of the cold fluid. The second heat dissipation strategy adds the pulse valve of the jet injector to the first heat dissipation strategy, adjusting it according to a preset frequency, and spraying it obliquely onto the chip surface through a micro-nozzle array; the high-speed pulse jet directly tears the thermal boundary layer around the chip, allowing the coolant to quickly absorb the chip's heat and convert it into a heat fluid.

9. A single-phase immersion liquid cooling heat dissipation control device, used to implement the single-phase immersion liquid cooling heat dissipation control method as described in any one of claims 1 to 7, characterized in that, Includes: V-shaped guide fins, coolant buffer zone inlet pipes, horn-shaped guide holes, first guide pipe, jet injector, cabinet, server chip, isolation board, and second guide pipe; The cabinet has V-shaped guide fins welded to both sides of the inner top. A coolant buffer zone inlet pipe is embedded in the center of the top of the cabinet. At least two jets are installed at the bottom of the coolant buffer zone inlet pipe. A first guide pipe is installed on one side of each jet. A horn-shaped guide hole is installed at the top of the first guide pipe. A server chip is placed in the middle between the first guide pipes. The horn-shaped guide hole is embedded in the isolation plate. Multiple second guide pipes are connected to the lower end of the isolation plate.

Images